Errors when randomizing Hexagonal Grid

Hi Everyone,

I am currently developing a program that generates a grid of regular hexagons, which I have been able to do

However, I am trying to make it so when a certain percentage is entered(third parameter in function), it will randomly delete that percentage of hexagon sides randomly in the figure that appears. However, I have been using the function randperm to generate random numbers corresponding to the index for which sides/corners to take out, and for some reason when I run it random lines are generated across the screen. Would anyone know why that is?

Thanks!

(the code attached might seem long, but a lot of it just repetitive)

[Lawrence Cheng]

Also, here is a picture of the output when it does mess up. For some reason it seems that works occasionally but messes up other times. Please let me know if you need more information!

[Mike Garrity]

I'm not clear on what this step is supposed to do:

    a(sample(1:k/2)) = NaN;
    b(sample(1:k/2)) = NaN;

If I stop there in the debugger, a and b are 1x7 arrays containing the coordinates of a single hexagon, while sample is a 1x700 full of large integers. That means that this step pads a & b way out with a mix of 0's and nan's. Then the next step samples them back down. The result, as you say, is to create a bunch of lines that go to zero. What were you trying to do here? I don't get your intent.

[Lawrence Cheng]

Hi Mike, Thanks for your response! For this program I am storing the individual x and y coordinate values into the array of a and b. The array sample is an array filled with random numbers made by randperm (these random numbers represent the indexes for the arrays a and b). Essentially, since I am trying to randomly delete random hexagonal sides, I am setting random x and y values to NaN so that in the output figure it does not show that particular side. For instance, the snippet you posted chooses the indexes from 1 to k/2 in the random array of sample, and thus the arrays a and b would have those specific elements set randomly to NaN.

[Lawrence Cheng]

Actually, I have tried both RandPerm and Datasample for the file, but both still have the same effect. I am also open to suggestions for different ways to randomize this without using these functions and for not selecting the same random number twice.

[Lawrence Cheng]

Sorry for all the comments, but I discovered that the program is more likely to break and draw those random lines across the screen when a larger number of random numbers are drawn together at once. For example, doing a(sample(1:500)) will break every time, while a(sample(1:50)) runs smoothly everytime. Hope that helps.

[Mike Garrity]

"For instance, the snippet you posted chooses the indexes from 1 to k/2 in the random array of sample, and thus the arrays a and b would have those specific elements set randomly to NaN."

Yes, but sample is bigger than a and b. This means that the a & b arrays get stretched to the size of the sample array. When this happens, any values you haven't specified get filled with 0's. I think that these zeros are what you're seeing.

You can see the pattern in this simple example:

rng(0)
a = 1:7
a(randi(100,[1 5])) = nan

The first disp shows a being a 1x7, but the second disp shows it being big enough to hold the last nan. But it got padded with a lot of zeros.

[Lawrence Cheng]

Ok, I will look into that.

Thank you for your help!